Mark Lingonqvist

Why the fuck would i need to know if a data breach is AI driven? What use is that to me?

Interfacing thought

one man's rubbish is another's data-driven design decision

Disp work for heavy underground machinery. In this case. a minetruck

if you insist on using AI to design stuff, reduce letter spacing if it is the only thing you do.

Alltså, hur fint är inte det här? UX-chefen på Saab ville få tag på mig, men efter en systemmigrering hade hon tappat bort min mejladress. Min hemsida var dessutom blockerad av Saab. Så hon letade upp min gamla Figma-portfolio och lämnade ett meddelande där istället. Man blir ju lite smickrad. Och imponerad av kreativiteten.

What my figma file sees when I manipulate it via MCP & Claude instead of touching the pixels myself like a peasant

well fuck how can i deauthorise computers i no longer have?

getting stuck back into gui and operator stations

claude's ui feel is a cut above vs code. something about the spacing and the use of separators makes you feel closer to the output even when asking them to manipulate material in Figma.

after listening to a cognitive ergonomicist's thesis presentation, the idea of fan-out number is starting to feel like a central point for a fleet overview GUI im making for control rooms. One operator, or team, looking after many autonomous units. how many units can one operator can run before performance drops? There's some research on how many robots one person can supervise - there's almost none on what the overview screen should look like tho. No tested layouts, no real proven patterns. using Oblivion drone grid as a fun reference - you can see which units are offline (not obviously enough imo). crucially, it doesn't show why, how long it's been down, or which one may get knocked out next. I'm working with the concepts of perceive, understand, project (Endsley's three levels of situation awareness). In my autonomous fleet overview, I'm playing with tiles that sort by urgency, healthy units that fade back, and tiles that carry state, reason and time to the next threshold. Each unit shows its intent and confidence. In team rooms every alert shows an owner, and handovers go unit by unit. Grouping by operator, drone health or geography might impact all three. Grouping by health makes failures easy to spot, but it splits up drones that share a task or area. Grouping by geography shows coverage gaps and lets you see when a problem is spreading across one zone. Grouping by operator shows who owns what, and whose workload is about to spike. I'm testing whether operators can switch between groupings without losing their place.

Just had a disastrous meeting with a senior stakeholder. It’s not all rainbows, that’s for sure.

Colleague is doing a 3D render of some GUIs/Interfaces I helped design. A teleremote station for subterranean fleets: two joysticks, a touchscreen, an e-stop, and two dozen autonomous machines run from an office far from the danger, dust and noise.

Control Room Design Guidelines: first draft. A colleague told me to "write these as if a junior designer or a new join were looking to these for guidance".